Update context and output limits for Vultr models

Adjusted the 'context' and 'output' token limits in TOML configs for deepseek-r1-distill-llama-70b, deepseek-r1-distill-qwen-32b, gpt-oss-120b, kimi-k2-instruct, and qwen2.5-coder-32b-instruct models to reflect new capacity constraints.

Author: cyberofficial

providers/vultr/models/deepseek-r1-distill-llama-70b.toml

@@ -13,8 +13,8 @@ input = 0.20
 output = 0.20
 
 [limit]
-context = 128_000
-output = 32_768
+context = 121_808
+output = 8_192
 
 [modalities]
 input = ["text"]

providers/vultr/models/deepseek-r1-distill-qwen-32b.toml

@@ -13,8 +13,8 @@ input = 0.20
 output = 0.20
 
 [limit]
-context = 128_000
-output = 32_768
+context = 121_808
+output = 8_192
 
 [modalities]
 input = ["text"]

providers/vultr/models/gpt-oss-120b.toml

@@ -13,9 +13,9 @@ input = 0.20
 output = 0.20
 
 [limit]
-context = 128_000
-output = 131_072
+context = 121_808
+output = 8_192
 
 [modalities]
 input = ["text"]
-output = ["text"]
+output = ["text"]

providers/vultr/models/kimi-k2-instruct.toml

@@ -13,9 +13,9 @@ input = 0.20
 output = 0.20
 
 [limit]
-context = 128_000
-output = 16_384
+context = 58_904
+output = 4_096
 
 [modalities]
 input = ["text"]
-output = ["text"]
+output = ["text"]

providers/vultr/models/qwen2.5-coder-32b-instruct.toml

@@ -13,9 +13,9 @@ input = 0.20
 output = 0.20
 
 [limit]
-context = 128_000
-output = 32_768
+context = 12_952
+output = 2_048
 
 [modalities]
 input = ["text"]
-output = ["text"]
+output = ["text"]